Researchers from the University of Waterloo Center for Automotive Research (WatCAR) in Canada are modifying a Lincoln MKZ Hybrid to autonomous drive-by-wire operation. The research platform, dubbed “Autonomoose” is equipped with a full suite of radar, sonar, lidar, inertial and vision sensors; NVIDIA DRIVE PX 2 AI platform (earlier post) to run a complete autonomous driving system, integrating sensor fusion, path planning, and motion control software; and a custom autonomy software stack being developed at Waterloo as part of the research.
Recently, the Autonomoose autonomously drove a crew of Ontario Ministry of Transportation officials to the podium of a launch event to introduce the first car approved to hit the roads under the province’s automated vehicle pilot program.
Operating at 24 trillion deep learning operations per second, DRIVE PX 2 enables Autonomoose to navigate Ontario’s city streets and highways, even in inclement weather.
The WatCAR research team has Autonomoose operating at level 2 autonomy, where the driver must be prepared to take over from the system in the event it fails to respond to a situation properly. Over the duration of the research program, they will advance the automation through level 3—where drivers can turn their attention away in certain environments, such as freeways—and ultimately reach level 4, where the automated system can control the car under most all circumstances.
|Picture: WatCAR. Click to enlarge.|
Ontario is the first province in Canada to create a pilot program to test automated vehicles on its roads. WatCAR was the first applicant and the first approved participant to test a vehicle on public roads. Public road testing of Autonomoose in both ideal and adverse weather conditions will begin early next year.
The province places no restriction on where these test vehicles can be driven—an advantage compared to most programs around the world, which restrict driving to certain areas of cities or highways.
Canada’s Natural Sciences and Engineering Research Council (NSERC) provided initial research funding for Autonomoose. Nine professors are involved from the Faculty of Engineering and Faculty of Mathematics. Specific projects include:
Simultaneous localization and mapping (SLAM) in all weather conditions.
Autonomous maneuvers under extreme conditions.
Power-management controllers for autonomous driving.
Feature-oriented engineering (FOE). Autonomous driving requires a multitude of distinct behaviors, called features (e.g. Lane Changing, Self-Parking). Feature-oriented engineering (FOE) promotes independent development of the features and ultimately feature interactions (FIs). FIs are a fundamental challenge in FOE.
Optimize self-driving for fuel efficiency and reduced emissions.
Runtime monitoring and reconfiguration infrastructure for autonomous driving.
Fault-tolerant electric/electronic (E/E) architectures for autonomous vehicles.
Functional safety for software and components of autonomous vehicle systems.